Yujie Zhu , Jiang Wei Wang , Yang Liu , Xiaohua Liu , Akihiro Kushima , Yihang Liu , Yunhua Xu , Scott X. Mao , Ju Li , * Chunsheng Wang , * and Jian Yu Huang *
In Situ Atomic-Scale Imaging of Phase Boundary Migration in FePO 4 Microparticles During Electrochemical Lithiation
ION
Orthorhombic Li x FePO 4 (0 ≤ x ≤ 1) system has attracted much attention for its application as a high power cathode material in lithium ion batteries. [ 1 ] Although the performance of this material has been greatly improved by cation doping, surface coating and size reduction, [ 2–4 ] the fundamental phase transfor-mation mechanisms accompanying lithiation/delithiation are still controversial. [ 5–19 ] Theoretical calculations have revealed that the lithiation/delithiation in LiFePO 4 is highly anisotropic with lithium ion diffusion being mainly confi ned into the chan-nels along the b -axis (space group Pnma ). [ 20–22 ] Ex situ high resolution transmission electron microscopy (HRTEM) study of microsized LiFePO 4 demonstrated that the phase boundary (PB) between LiFePO 4 and FePO 4 tends to lie on the bc plane. [ 23 ] Several variants of the PB migration mechanism have been proposed, such as the “domino-cascade” model and the “new core-shell” model. [ 10 , 11 ] Singh et al. developed a compre-hensive surface-reaction-limited intercalation theory and fi rst predicted that the LiFePO 4 /FePO 4 PB migrated perpendicular to the lithium ion insertion direction like a “travelling wave”. [ 12 ] Recently, Tang et al . proposed an overpotential-dependent phase transformation model by considering both the ani-sotropic lithiation and misfi t strain properties, [ 6 , 17 ] where crystal → amorphous transitions occur at intermediate overpo-tential, and crystal → crystal transitions occur at low as well as high overpotentials. [ 7 ] Under large overpotentials, the PB of FePO 4 /LiFePO 4 was predicted to align on the ac plane instead of bc plane. [ 13 ] The solid solution mechanism was also proposed
Y. Zhu, [+] Y. Liu, Dr. Y. Xu, Prof. C. WangDepartment of Chemical and Biomolecular Engineering University of Maryland College Park, Maryland 20742, USA E-mail: [email protected] J. W. Wang, [+] Prof. S. X. MaoDepartment of Mechanical Engineering and Materials Science University of Pittsburgh Pittsburgh, Pennsylvania 15261, USA Dr. Y. Liu, [+] Dr. X. Liu, Dr. J. Y. HuangCenter for Integrated Nanotechnologies Sandia National Laboratories Albuquerque, New Mexico 87185, USA Email: [email protected] Dr. A. Kushima, Prof. J. LiDepartment of Nuclear Science and Engineering and Department of Materials Science and Engineering Massachusetts Institute of TechnologyCambridge, Massachusetts 02139, USAE-mail: [email protected] [+] These authors contributed equally to this work.
DOI: 10.1002/adma.201301374
Adv. Mater. 2013, 25, 5461–5466
for nano-LiFePO 4 , [ 14–16 ] which suggested a single-phase lithi-ation pathway rather than nucleation and growth of a second phase during lithiation/delithiation. Bai et al. quantitatively demonstrated that the expected phase transformation would be suppressed in nano-LiFePO 4 and the lithiation would follow a solid solution behavior if the discharge current exceeded a critical value. [ 15 ] However, the single-phase solid solution transi-tion pathway usually is expected in nanosized particles ( ∼ 40 nm diameter with sloping charge/discharge voltage curves), [ 16 ] and not in microsized particles.
Obviously, real-time atomic-scale observation of lithiation/delithiation of LiFePO 4 is critical to clarify the phase transfor-mation mechanisms. However, such dynamic observations have not been achieved due to technical diffi culties associated with the experiments. Here we report the fi rst dynamic HRTEM observations of the PB migration in microsized FePO 4 single-crystals during electrochemical lithiation, providing the fi rst direct atomic-scale evidence for the PB migration mechanism.
In situ HRTEM observation of electrochemical lithiation in FePO 4 was conducted in an all-solid electrochemical cell setup, [ 24 , 25 ] which consisted of a FePO 4 crystal working elec-trode, a naturally-grown Li 2 O solid electrolyte, and a bulk lithium metal counter electrode ( Figure 1 a). The orthorhombic FePO 4 was obtained from chemical delithiation of LiFePO 4 which was synthesized from a hydrothermal method. [ 26 ] The crystal structures of both LiFePO 4 and FePO 4 were
Figure 1 . In situ TEM electrochemical experiment setup and morphology of FePO 4 crystal. (a) Schematic illustration of the in situ electrochemical cell. (b) A pristine FePO 4 crystal in [ ̄100 ] zone axis was connected with the Li 2 O electrolyte to form an electrochemical device. The propagation of a phase boundary between FePO 4 and LiFePO 4 in the rectangle zone marked by yellow dashed line was captured in real time under HRTEM, as presented in Figure 2 . (c) The electron diffraction pattern (EDP) of the FePO 4 crystal as shown in (b), indicating it is a single-crystal with the orthorhombic structure.
Figure 2 . Migration of phase boundary between FePO 4 and LiFePO 4 along the [010] direction during lithiation. (a) After re-turning on the electron beam, a clear phase boundary between FePO 4 and LiFePO 4 was already developed due to lithiation, as pointed out by the red arrows. A layer of 10 nm amorphous carbon (a-C) was deposited on the surface of FePO 4 crystal to decrease the electron beam induced damage. (b) At 176 s, the thickness of the LiFePO 4 layer increased. The phase boundary was pointed out by the red arrows. The invert “T” marks mis-match dislocations at the phase boundary. (c) Inverse FFT (IFFT) image of (b), showing the distribution of mismatch dislocations near the phase boundary. (d) The FFT patterns of FePO 4 and LiFePO 4 produced from FePO 4 and LiFePO 4 regions in (b). To make comparison, FFT pat-tern of LiFePO 4 (green) was overlaid with that of FePO 4 (yellow). (e,f) Lattice spacing of (002) plane from FePO 4 and LiFePO 4 was measured from HRTEM image shown in (b), showing the lattice spacing of (002) plane decreased about 1.7% after lithium ion insertion into FePO 4 .
characterized by a powder X-ray diffraction (XRD) (Supporting Information Figure S1). The morphology of FePO 4 was character-ized by a scanning electron microscopy (SEM) (Supporting Information Figure S2). The electrochemical behaviors of chemi-cally delithiated FePO 4 were measured in a traditional liquid electrolyte cell (Supporting Information Figure S3), which showed well-defi ned voltage plateaus during both charge and discharge processes similar to the elec-trochemical behaviors of typical LiFePO 4 electrodes, demonstrating that the lithiation mechanism of chemically delithiated FePO 4 also refl ects the lithiation mechanism of electrochemically delithiated FePO 4 . [ 27 ] More experimental details (materials synthesis, in situ experiment set up and data processing procedure) are provided in Experimental Sec-tion and Supporting Information.
Figure 1 b presents a TEM image of a pristine FePO 4 crystal connected to the Li 2 O electrolyte. The electron diffraction patterns (EDPs) of the FePO 4 crystal (Figure 1 c and Supporting Information Figure S4) indi-cated that it was a single-crystal with the orthorhombic structure, and the growth direction was [001] (Supporting Infor-mation Figure S4). Before lithiation, the FePO 4 crystal was tilted into [ ̄100 ] zone axis (Figure 1 c) and the maximum thickness of the particle along this axis was around 300 nm (Supporting Information Figure S5). Different from many anode materials (Si, Sn etc.), the volume change for FePO 4 upon lithiation is only 7%, which is too small to be observed at a low magnifi cation TEM image. So HRTEM was used to follow the lithiation process.
Figure 2 shows HRTEM images of the PB evolution, from the rectangular zone in Figure 1 b, during lithiation of FePO 4 . The thickness along a -axis for this HRTEM obser-vation region is less than 300 nm (Supporting Information Figure S5). A positive voltage of 2 V versus lithium metal was applied to the FePO 4 crystal in Figure 2 to drive lithium ion insertion into the crystal. To increase the conductivity and decrease the electron beam damage to FePO 4 , 10 nm amorphous carbon
(a-C) was coated on the surface of FePO 4 crystals (Figure 2 a) and weak electron beam was used for imaging. To further mini-mize the electron beam damage to the sample, the electron beam was blanked immediately after the voltage was applied, and was re-turned on 10 min later to take the HRTEM images. It was found that below the a-C coating layer, a thin layer of LiFePO 4 was already formed on the surface of FePO 4 due to lithium insertion (Figure 2 a). A clear PB (pointed out by the red arrows in Figure 2 a), as evidenced by the image contrast
and the misfi t strain near the PB, was developed between FePO 4 and LiFePO 4 due to lithium ion insertion. After 176 s, a thicker LiFePO 4 layer (pointed out by red arrows in Figure 2 b) was developed. The PB migrated from the surface inward the core as more lithium ions were inserted into the FePO 4 (Sup-porting Information Figure S6). From the displacement of PB (Supporting Information Figure S7), the corresponding dis-charge C-rate for the particle is roughly estimated to be 0.1C. The PB was clearly a sharp interface (Figure 2 a, b), which is
Figure 3 . Step-like phase boundary between FePO 4 and LiFePO 4 and its migration along the [010] direction during lithiation. (a) A HRTEM image of the pristine FePO 4 in [ ̄100 ] zone axis. The crystal orientation and (020) plane were denoted. (b) At 215 s after applying the voltage, a step-like phase boundary was formed between FePO 4 and LiFePO 4 , as pointed out by the yellow dashed line. (c) At 282 s, the thickness of the LiFePO 4 layer increased as the step-like phase boundary propagating along the [010] direction. The regions, marked by white dashed lines, show where we took the FFT pat-terns for LiFePO 4 and FePO 4 phases. The inset image shows Figure 3 d. (d) Inset of (c) with inverse FFT (IFFT) image from the rectangle zone marked by red dashed line in (c), showing the lattice mismatch induced dislocations at the phase boundary between FePO 4 and LiFePO 4 .(e) The FFT pattern of FePO 4 side showing sharp single spots at each diffraction points. The inset image shows that each diffraction spots do have one point. (f) The FFT pattern of the LiFePO 4 /FePO 4 co-existing zone. Two sets of diffraction patterns are identifi ed to be LiFePO 4 and FePO 4 , respectively. The inset image shows that each diffraction spots do split into two points. (g,h) Lattice spacing of (020) plane for FePO 4 and LiFePO 4 .
Adv. Mater. 2013, 25, 5461–5466
consistent with the electron energy loss spectroscopy (EELS) study reported by Laffont et al . [ 11 ] in which the PB was shown to be the juxtaposition of FePO 4 and LiFePO 4 rather than a solid solution region with a concentration gradient. EELS char-acterization was performed on both the pristine FePO 4 and the LiFePO 4 generated by in situ lithiation for the particle in Figure 2 , and the results confi rmed formation of LiFePO 4 in FePO 4 (Supporting Information Figure S8 and corresponding text). [ 11 ] Detailed analyses of the HRTEM images (see data pro-cessing procedure for HRTEM images in Supporting Informa-tion) indicated that the PB was nearly parallel to the ac plane, and it migrated along the b -axis inward. It is worth noting that although the FePO 4 crystal was connected with Li 2 O solid elec-trolyte on the left edge where ab face was exposed (Figure 1 b), the lithium ion insertion did not occur along the c -axis but along the b -axis which is far away from the lithium source, indi-cating that lithium ion was inserted into FePO 4 only along the b -axis of FePO 4 crystal, probably after transported via surface diffusion. The one-dimensional lithiation mechanism along the [010] direction ( b -axis) is in accordance with theoretical calcula-tions. [ 20–22 ] Therefore, our in situ HRTEM observation provides the fi rst direct evidence for the anisotropic lithiation mecha-nism in FePO 4 .
The phase transformation from FePO 4 to LiFePO 4 induces nearly periodic array of dislocations on the FePO 4 side at the PB, as outlined by the inverted “T” in Figure 2 b,c. Figure 2 d pre-sents the superimposed Fast Fourier Transformation (FFT) pat-terns of FePO 4 (yellow dots) and LiFePO 4 (green dots) produced from the HRTEM image (Figure 2 b). Obviously, after lithiation, the orientation of newly formed LiFePO 4 slightly rotated com-paring with FePO 4 , which was also observed by Chen et al. [ 23 ] in chemically delithiated Li 0.5 FePO 4 using ex situ HRTEM. This may be attributed to inhomogeneous elastic deformation of the particle to accommodate the transformation strain. Also, as shown in Figure 2 e-f, after FePO 4 was transformed into LiFePO 4 , the lattice spacing of the (002) plane decreased about 1.7%, which is close to the theoretically expected lattice differ-ence between FePO 4 and LiFePO 4 in this direction, i.e. 1.9%. If the PB was fully coherent, the lattice constant near PB should be quite different from the stress-free lattice constant of the par-ticular phase, and should approach the mean of the two phases. The fact that the measured lattice constants are quite close to the stress-free lattice constants means majority of the elastic misfi t energy along this direction has indeed been relaxed away, due to the presence of the misfi t dislocations.
The PB migration mechanism is repeatable in our experi-ments. Figure 3 are the HRTEM images showing the PB migration in another FePO 4 single-crystal during lithiation. A positive voltage of 2.7 V versus lithium metal was applied to the FePO 4 crystal. The pristine FePO 4 crystal had a uniform thick-ness (Supporting Information Figure S9) and was coated with a thin layer of a-C (Figure 3 a). After lithiation for 215 s, a PB on the (020) plane with few steps was formed between FePO 4 and LiFePO 4 , as marked by the yellow dashed line in Figure 3 b, and the lithium ion insertion direction was also [010] direc-tion ( b -axis). The step-like PB might be caused by the lithium ion concentration gradient along the a-C thin layer, where the right portion of the particle had higher lithium ion concentra-tion because it was closer to the lithium source (Supporting
5463wileyonlinelibrary.combH & Co. KGaA, Weinheim
5464
www.advmat.dewww.MaterialsViews.com
CO
MM
UN
ICATI
ON
Information Figure S9). After another 67 s, the thickness of the
LiFePO 4 layer increased as the step-like PB propagating along the [010] direction (Figure 3 c), and it was found that these steps moved along the [010] direction by simply following the move-ment of PB. Figure 3 d (inset of Figure 3 c) presents the inverse FFT (IFFT) image from the rectangle area, marked by red dashed line in Figure 3 c, showing the mismatch dislocations at the PB, which is similar to the results shown in Figure 2 c. The FFT patterns, from both the single-phase FePO 4 and the two-phase FePO 4 /LiFePO 4 regions marked by white dashed squares in Figure 3 c, were shown in Figure 3 e,f, respectively. For the FePO 4 region, very sharp single diffraction spots were observed in the pattern (Figure 3 e and its inset). However, the diffraction spots in the FFT pattern from the two-phase LiFePO 4 /FePO 4 region were split, showing two sets of FFT patterns (Figure 3 f and its inset) and diffuse intensity distributions, which is obviously different from the single-phase pattern shown in Figure 3 e. By measuring the lattice spacing in Figure 3 f, one set of the FFT pattern was identifi ed as LiFePO 4 , while the other was from FePO 4 . As shown in Figure 3 g,h, after FePO 4 was transformed into LiFePO 4 , the lattice spacing of the (020) plane increased by about 3.8%, which is consistent with the theoretical lattice misfi t value (i.e. 3.6%) in this plane. Besides the two particles presented in Figures 2 and 3 , the third FePO 4 particle was also studied and showed the similar results (Sup-porting Information Figure S10). These similar results further confi rmed that the dynamic PB in the FePO 4 microparticles was parallel to the (010) plane, and its propagation direction was along [010].
The theoretical misfi t strains along a , b and c directions between LiFePO 4 and FePO 4 are δ [100] = 5%, δ [010] = 3.6%, and δ [001] = − 1.9%, [ 28 ] respectively. It is obvious that a -axis misfi t contributes the most to the elastic misfi t strain, which unfor-tunately we were not able to observe due to the zone axis con-dition in imaging. However, the dislocation nucleation is so favorable, as shown in Figures 2 and 3 , that at room tempera-ture even a small misfi t strain of 1.9% can drive it. This implies that the larger misfi t strain of 5.0% in a -axis, which ought to give larger driving force for dislocation nucleation, should sus-tain signifi cant coherency loss effi ciently as well. In Figure 2 b, we count 5 misfi t dislocations over a length of ∼ 45 nm along the c -axis. The Burgers vector was identifi ed to be [011]/2 (Sup-porting Information Figure S10 d). So the contribution to c -axis misfi t displacement per dislocation was 4.7 Å/2 = 2.35 Å, and fi ve dislocations would contribute 2.35Å × 5 = 11.75 Å of ine-lastic accommodation in c- axis with 11.75 Å/ 45 nm = 2.6%, which is very close to the theoretical mismatch of 1.9% in c -axis, given the statistical sampling errors in a small view fi eld. Thus, the system is able to generate just the right density of dis-locations to nearly entirely cancel out the misfi t elasticity strain in c -axis. Dislocation nucleation was apparently facile in this system, which enables the coherency loss. Dislocation migra-tion was also sustainable, since the dislocations appeared to follow the motion of the chemical interface (Supporting Infor-mation Figure S6).
Previous theoretical models assumed that the differences in lattice constant between LiFePO 4 and FePO 4 are entirely taken up as elastic strain energy when phase transformation hap-pens, [ 6 , 10 ] giving rise to local stress as high as ∼ 1 GPa. Such a
high stress may couple into lithium migration and give rise to interesting physical effects such as exceedingly fast PB migra-tion in nanosized LiFePO 4 . [ 10 ] However, our in situ experiment reveals that, at least for particles with facile generation of misfi t dislocations, the elastic strain energy term should be greatly reduced. Facile generation of misfi t dislocations should favor the particle to be a two-phase mixture with a sharp interface between Li 1- x FePO 4 and Li x FePO 4 ( x ∼ 0) phases, rather than as a single-phase solid solution (no solubility gap), since the elastic coherency energy closes up the spinodal gap and delays or prevents the onset of concentration wave instability of super-saturated homogenous solid solution. [ 15 ] While the coherency loss caused by misfi t dislocations favors the LiFePO 4 /FePO 4 two-phase transition mechanism, it also has a signifi cant effect on the PB orientation. As recently suggested by Cogswell et al. [ 29 ] through a fully anisotropic analysis, a fully coherent PB between LiFePO 4 and FePO 4 would align on the (101) plane which was shown to be the lowest energy orientation. The observed (100) PB by ex situ HRTEM was attributed to a par-tial loss of coherency along c -axis. [ 23 ] In our study, the observed (010) PB is a consequence of even more severe coherency loss, since we clearly observed coherency loss happened along both b- and c- axis, and it was also highly favorable along a- axis as discussed above. Different strain relaxation mechanism along a- axis might alter the picture of observed phase boundary ori-entation because the lattice misfi t is the largest along this axis. Also, in reference to Tang et al . 's lithiation model, [ 13 , 17 ] while a high overpotential may override the thermodynamic considera-tion of strain energy and rotates the (100) inclined PB towards kinetics-controlled (010) inclination, introducing the misfi t dis-locations would serve a similar purpose, by reducing the strain energy itself.
While the introduction of misfi t dislocations has signifi cant impacts on the thermodynamics of Li x FePO 4 (0 < x < 1) system, it also infl uences the kinetic processes of phase transformation. First, the coherent PB to semicoherent PB transition is itself a thermally activated, size-dependent process with possibility of hysteresis, and depends sensitively on “extrinsic” or processing-dependent factors such as crystalline quality of the particle, confi nement of nearby particles and binder, etc., which are not necessarily just functions of the particle size. Second, while the presence of misfi t dislocations reduces the total energy of system, they also “pin” the chemical interface, in the sense that they need to co-move with the advancing chemical front by climbing, which is another thermally activated dissipative process that is generally quite slower than a coherent displacive process. [ 10 ]
Figure 4 shows the dynamic lithiation mechanism discov-ered in this study. With lithium ion insertion only along the [010] direction, the reduction of lattice misfi t strain energy by dislocations favors the two-phase transition mechanism and changes the PB from the thermodynamics-controlled (100) plane (or (101) plane) to kinetics-controlled (010) plane. [ 29 ] Our in situ HRTEM observations provide evidence that microscopic damage in the form of nucleated dislocations accompanies chemical transformations in microparticles. These dislocations are expected to accumulate during electrode cycling, which could form preferential sites for cracking. This is expected to be a general degradation mechanism in Li intercalation
Figure 4 . Schematic illustrations of lithiation mechanism discovered in this study. The migration direction of the phase boundary is the same as the lithium ion diffusion direction (i.e. [010] direction), which was observed in this study (Figures 2 and 3 ). The green arrows mark the lithium ion insertion direction.
compounds. Here, it needs to be emphasized that the FePO 4 particles studied here are microsized, so the lithiation mecha-nism is expected to change when the sample goes down to nanoscale as suggested by recent studies. [ 28–31 ] It has been shown that particle size, as well as electrode nanostructures (such as porous electrode etc.), plays a critical role on deter-mining the phase behavior of Li x FePO 4 system. [ 14–16 , 27–32 ] For example, Weichert et al. [ 27 ] recently studied the PB propagation in a LiFePO 4 particle, whose size was much larger than the samples in present study, upon chemical delithiation by in situ optical microscopy and the PB was shown to migrate along the [001] direction ( c -axis). which was clearly different from the pre-sent and previous HRTEM results. [ 23 ] Also, signifi cant amount of cracks were observed and the formed FePO 4 was highly porous. The transformation was shown to be diffusion-limited due to the large length scale. Decreasing the sample size to nanoscale would result in a different regime. Besides the short ion transport length, it has been shown that the lithium ion
diffusivity in nanoparticles is much higher than that in the bulk materials. [ 30 ] Also, it would be more diffi cult for dislocations to nucleate in smaller particles, [ 33 ] and a coherent PB is preferable in nanoparticles. This is supported by previous study in which the 43-nm LiFePO 4 particle was shown to have a much higher retained strain than the 113-nm sample when two samples were at the same state-of-charge. [ 28 ] The coherency strain also changed the phase diagram in Li x FePO 4 system, [ 29 ] specifi cally, stabilizing the solid solution at relatively low temperature and shrinking the miscibility gap between FePO 4 and LiFePO 4 .
Cycle life and capacity fading of batteries are closely related to the fatigue of electrodes, which is caused by accumulation of microscopic damage defects such as dislocations during elec-trochemical cycling, that eventually leads to cracking. [ 23 , 27 ] We have identifi ed a detailed pathway where dislocations may be generated, due to coherency loss transition during electrochem-ical lithiation. Since coherency loss is more likely in microsized particles than in nanosized particles, from well-known size effect similar to that of epitaxial thin fi lm growth and alloy pre-cipitation, we can predict that nanosized particles should have better fatigue life. Also, the accumulation of dislocations may lead to amorphization, which was observed in many electrode materials during electrochemical cycling. [ 7 ] Recently, solid-state amorphization due to accumulation of dislocations has been directly observed. [ 34 , 35 ]
In summary, we report the fi rst real-time atomic-scale observation of the phase boundary migration mechanism and the anisotropic lithiation in FePO 4 microparticles. The phase boundary was shown to align along the (010) plane and move towards the [010] direction which is the same as the lithium ion diffusion direction. This result was shown to be caused by relaxation of majority of the elastic strain energy, as demon-strated by the observed periodic dislocations along the phase boundary and the measured lattice spacings which are quite close to the stress-free lattice constants of FePO 4 and LiFePO 4 . Our in situ observation provided clear evidence for the phase boundary migration mechanism in microsized FePO 4 /LiFePO 4 system. The results reported here constitute the fi rst real-time atomic-scale experimental evidence of the phase transforma-tion mechanism in FePO 4 /LiFePO 4 system, which is currently under intensive experimental and theoretical investigations.
Experimental Section Materials synthesis and characterization : LiFePO 4 was synthesized
using a hydrothermal method. [ 26 ] LiOH aqueous solution (10 mL of 4 M) was mixed with aqueous solution (5 mL) of H 3 PO 4 (0.015 mol) and (NH 4 ) 2 HPO 4 (0.005 mol) to form a white suspension. Then, FeSO 4 · 7H 2 O aqueous solution (10 mL of 2 M) was slowly added into above suspension with continuous stirring and argon purging. The molar ratio of Li:Fe:P was kept at 2:1:1. The mixture was transferred to a Parr autoclave, which was then held at 180 ° C for 12 h. After natural cooling to room temperature, the product was collected and washed with ethanol and deionized (DI) water for several times. The fi nal product was dried at 80 ° C in a vacuum oven for overnight.
FePO 4 was obtained from chemical delithiation of LiFePO 4 using nitronium tetrafl uoro-borate NO 2 BF 4 in acetonitrile. LiFePO 4 (0.1 g) was added into a solution of NO 2 BF 4 (0.17 g) in acetonitrile (10 mL). The mixture was stirred for 24 h at room temperature with continuous Ar bubbling, followed by centrifugation and washing with acetonitrile and
5465wileyonlinelibrary.combH & Co. KGaA, Weinheim
5466
www.advmat.dewww.MaterialsViews.com
CO
MM
UN
ICATI
ON
[ 1 ] A. K Padhi , K. S. Nanjundaswamy , J. B. Goodenough , J. Electrochem. Soc. 1997 , 144 , 1188 .
[ 2 ] S. Y. Chung , J. T. Bloking , Y. M. Chiang , Nature Mater. 2002 , 1 , 123 [ 3 ] B. Kang , G. Ceder , Nature 2009 , 458 , 190 .
DI water for several times. The product was dried at 80 ° C in a vacuum oven for overnight. The crystal structures of both LiFePO 4 and FePO 4 were characterized by a powder X-ray diffraction (XRD) (Supporting Information Figure S1). Both XRD patterns were indexed in the orthorhombic ( Pnma ) crystallographic system (Supporting Information Figure S1) and showed that there were no detectable impurities in the samples. The morphology of obtained FePO 4 was characterized by SEM. The results are shown in Figure S2. The electrochemical responses of FePO 4 were tested and shown in Figure S3. The orientation of FePO 4 crystals was determined by HRTEM and the results were shown in Figure S4. Figure S5 shows a double tilt experiment for a typical FePO 4 particle.
Set up of the in situ HRTEM experiment : For the in situ HRTEM experiment, the FePO 4 crystals were glued onto the aluminum rod with conductive epoxy. Fresh lithium metal was scratched from a fresh cut lithium metal surface by using a tungsten wire inside a glove box, and transferred into the TEM using a sealed bag fi lled with dry helium. During the transfer process, the lithium metal was exposed in air for about 2 s. The naturally-grown Li 2 O served as the solid electrolyte. It has been shown that Li 2 O can be functioned as an effective solid electrolyte for the lithium ion transport. [ 24 , 25 ] The Li/Li 2 O was driven to contact the FePO 4 crystal using a piezomanipulator (Nanofactory transmission electron microscopy–scanning tunneling microscopy (TEM-STM) holder). Comparing with the conventional liquid cell, this kind of solid cell offers the advantage of direct observation of the microstructure evolution inside TEM, especially for the in situ HRTEM.
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements This work is supported as part of the Nanostructures for Electrical Energy Storage, an Energy Frontier Research Center funded by the U.S. Department of Energy, Offi ce of Science, and Offi ce of Basic Energy Sciences under Award Number DESC0001160. Financial support in part by the National Science Foundation under Contract No. CBET0933228 (Dr. Maria Burka, Program Director) is gratefully acknowledged. A. K. and J. L. acknowledge support by NSF DMR-1008104 and DMR-1120901. The authors acknowledge the technical support of the NanoCenter in University of Maryland. Portions of this work were supported by a Laboratory Directed Research and Development (LDRD) project at Sandia National Laboratories (SNL). The LDRD supported the development and fabrication of platforms. The NEES center supported the development of TEM techniques. CINT supported the TEM capability. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
[ 4 ] X. L. Wu , L. Y. Jiang , F. F. Cao , Y. G. Guo , L. J. Wan , Adv. Mater. 2009 , 21 , 2710 .
[ 5 ] C. V. Ramana , A. Mauger , F. Gendron , C. M. Julien , K. Zaghib , J. Power Sources 2009 , 187 , 555 .
[ 6 ] M Tang , J. F. Belak , M. R. Dorr , J. Phys. Chem. C 2011 , 115 , 4922 . [ 7 ] Y.-H. Kao , M. Tang , N. Meethong , J. Bai , W. C. Carter , Y.-M. Chiang ,
Chem. Mat. 2010 , 22 , 5845 . [ 8 ] V. Srinivasan , J. Newman , J. Electrochem. Soc. 2004 , 151 , A1517 . [ 9 ] A. S. Andersson , J. O. Thomas , J. Power Sources 2001 , 97-8 , 498 . [ 10 ] C. Delmas , M. Maccario , L. Croguennec , F. Le Cras , F. Weill , Nature
Mater. 2008 , 7 , 665 . [ 11 ] L. Laffont , C. Delacourt , P. Gibot , M. Y. Wu , P. Kooyman ,
C. Masquelier , J. M. Tarascon , Chem. Mater. 2006 , 18 , 5520 . [ 12 ] G. K. Singh , G. Ceder , M. Z. Bazant , Electrochim. Acta 2008 , 53 ,
7599 . [ 13 ] Y. M. Chiang , Science 2010 , 330 , 1485 . [ 14 ] R. Malik , F. Zhou , G. Ceder , Nature Mater. 2011 , 10 , 587 . [ 15 ] P. Bai , D. A. Cogswell , M. Z. Bazant , Nano Lett. 2011 , 11 , 4890 . [ 16 ] P. Gibot , M. Casas-Cabanas , L. Laffont , S. Levasseur , P. Carlach ,
S. Hamelet , J.-M. Tarascon , C. Masquelier , Nature Mater. 2008 , 7 , 741 .
[ 17 ] M. Tang , W. C. Carter , J. F. Belak , Y. M. Chiang , Electrochim. Acta 2010 , 56 , 969 .
[ 18 ] W. Dreyer , J. Jamnik , C. Guhlke , R. Huth , J. Moskon , M. Gaberscek , Nature Mater. 2010 , 9 , 448 .
[ 19 ] L. Gu , C. Zhu , H. Li , Y. Yu , C. Li , S. Tsukimoto , J. Maier , Y. Ikuhara , J. Am. Chem. Soc. 2011 , 133 , 4661 .
[ 20 ] D. Morgan , A. Van der Ven , A. G. Ceder , Electrochem. Solid State Lett. 2004 , 7 , A30 .
[ 21 ] M. S. Islam , D. J. Driscoll , C. A. J. Fisher , P. R. Slater , Chem. Mater. 2005 , 17 , 5085 .
[ 22 ] J. L. Allen , T. R. Jow , J. Wolfenstine , Chem. Mater. 2007 , 19 , 2108 .
[ 23 ] G. Y. Chen , X. Y. Song , T. J. Richardson , Electrochem. Solid State Lett. 2006 , 9 , A295 .
[ 24 ] X. H. Liu , H. Zheng , L. Zhong , S. Huang , K. Karki , L. Q. Zhang , Y. Liu , A. Kushima , W. T. Liang , J. W. Wang , J.-H. Cho , E. Epstein , S. A. Dayeh , S. T. Picraux , T. Zhu , J. Li , J. P. Sullivan , J. Cumings , C. Wang , S. X. Mao , Z. Z. Ye , S. Zhang , J. Y. Huang , Nano Lett. 2011 , 11 , 3312 .
[ 25 ] X. H. Liu , Y. Liu , A. Kushima , S. Zhang , T. Zhu , J. Li , J. Y. Huang , Adv. Energy Mater. 2012 , 2 , 722 .
[ 26 ] K. Dokko , S. Koizumi , H. Nakano , K. Kanamura , J. Mater. Chem. 2007 , 17 , 4803 .
[ 27 ] K. Weichert , W. Sigle , P. A. Van Aken , J. Jamnik , C. Zhu , R. Amin , T. Acarturk , U. Starke , J. Maier , J. Am. Chem. Soc. 2012 , 134 , 2988 .
[ 28 ] N. Meethong , H. Y. S. Huang , S. A. Speakman , W. C. Carter , Y. M. Chiang , Adv. Funct. Mater. 2007 , 17 , 1115 .
[ 29 ] D. A. Cogswell , M. Z. Bazant , ACS Nano 2012 , 6 , 2215 . [ 30 ] R. Malik , D. Burch , M. Z. Bazant , G. Ceder , Nano Lett. 2010 , 10 ,
4123 . [ 31 ] M. Z. Bazant , Phase-fi eld theory of ion intercalation kinetics,
arXiv:1208.1587v1. [ 32 ] T. R. Ferguson , M. Z. Bazant , J. Electrochem. Soc. 2012 , 159 ,
A1967 . [ 33 ] T. Zhu , J. Li , Prog. Mater. Sci. 2010 , 55 , 710 . [ 34 ] J. Y. Huang , L. Zhong , C. M. Wang , J. P. Sullivan , W. Xu , L. Q. Zhang ,
S. X. Mao , N. S. Hudak , X. H. Liu , A. Subramanian , H. Fan , L. Qi , A. Kushima , J. Li , Science 2010 , 330 , 1515 .
[ 35 ] S. W. Nam , H. S. Chung , L. C. Yu , Q. Liang , J. Li , Y. Lu , A. T. C. Johnson , Y. Jung , P. Nukala , R. Agarwal , Science 2012 , 336 , 1561 .
